Darren Dahly

@statsepi.bsky.social

13551 Followers 2618 Following

Epidemiologist + Statistician | Clinical Research Facility - University College Cork | UCC School of Public Health | #ClinicalTrials #Epidemiology #Statistics #RStats Posts not for everyone ¯\_(ツ)_/¯ Views mine -> https://statsepi.substack.com/

Within academic institutions, actively carrying out faculty searches, or even establishing endowed chairs, on research topics that might be of greater interest to conservatives may provide a critical path forward
At the recent launch of the new AI policy at Ireland’s largest university, the Microsoft man representing the UCD-industry partnership told those in attendance that women in Ireland are not adopting AI tools as fast as men. He then said if he had a magic wand to grant one wish, he would get the Government to supercharge Ireland’s AIready training program so that more women would get on board and start using AI. Rather than uncritically accepting the Microsoft executive’s suggestion that big tech companies want to collaborate with government and universities to help women, what if the problem is not the women but big tech and their AI platforms?
miss_strings <- c( "No data", "NA", "Not done", "no data", "not done", "Not checked", "not avail", "Not performed", "Not commented on", "Nor commented", "Not commented", "Not avail" )
2 plots with years along the x-axis and % change on the y axis with sin-wave like lines running across both. Each plot also has a set of vertical lines with numeric labels. Relevant label code for the bottom plot: geom_text_repel( aes(x = date, y = 0.16, label = code), size = 3, direction = "both", color = cols_[3], ylim = c(0.16, 0.30), min.segment.length = 0, max.overlaps = Inf, hjust = 0, nudge_x = 1, nudge_y = 1, segment.alpha = 0.3 ) + coord_cartesian(ylim = c(-0.15, 0.15), clip = "off")
Real Clear’s evolution traces a similar path as other right-leaning political news outlets that have adapted to the upheaval of the Trump era by aligning themselves with the president and his large following, its writers taking on his battles and raging against the left. As the administration lurched from one crisis after another — impeachment, the coronavirus, a lost election the president refuses to concede — Real Clear became one of the most prominent platforms for elevating unverified and reckless stories about the president’s political opponents, through a mix of its own content and articles from across conservative media. In recent days, as Mr. Trump and his loyalists repeated baseless claims of rampant voter fraud and counting errors, Real Clear Politics gave top billing to stories that reinforced the false narrative that the president could still somehow eke out a win. Headlines on Monday — more than a week after Mr. Biden had clinched the race — included “There’s Good Reason Not To Trust Election Results” and “Trump Attorney Says Results in Several States Will Be Overturned.”
As a small state, Ireland has historically benefited from rules-based international institutions. We do not possess the military power to impose our will on others. We rely instead on diplomacy, international law, and multilateral cooperation. If those institutions are failing, surely our first instinct should be to strengthen and reform them rather than quietly abandon them. Ireland has a unique credibility on this issue. Few countries can claim our record in peacekeeping. Irish soldiers have served with distinction under the blue helmet for generations. We have consistently punched above our weight in diplomacy and conflict resolution. We should be leading conversations about reforming the United Nations, particularly the Security Council veto, rather than treating the organisation as an inconvenience to be bypassed. That would be a far more ambitious vision of Irish foreign policy than simply aligning ourselves ever more closely with the strategic priorities of larger powers.
Care is therefore needed in how TTEs are reported, to prevent any unintended obfuscation of the study limitations. We have already seen some misleading language in TTEs. While peer-reviewing an emulated trial in ART, we saw that investigators had included the subheading “Randomization”. In this section of the manuscript, they described how participants “were randomly assigned” to treatment or control groups. This was a retrospective, observational study – no one had been randomly assigned, but a reader could not be blamed for thinking otherwise.
Again, I am not proposing that preregistration be banned; rather, I am suggesting that it not be required because it is neither necessary nor sufficient for the production of high-quality science. Anyone who doubts this might note that the United States recently launched a $10 billion experiment with the James Webb Space Telescope designed to look into the far reaches of the universe without a single preregistered hypothesis. Rather, the telescope was designed for pure discovery, without pre-established beliefs about what might be found. Furthermore, other tools remain available to advance the goals that advocates of preregistration claim it offers. At the very least, preregistration does not provide the only mechanism to advance goals of transparency, replicability, and scientific honesty.
Not a single person showed up for my session today at a large district conference. This is what I learned… That the session time ended up being well spent sending self-deprecating selfies of an empty room behind me to colleagues, eating some gummy candies and catching up on email. That’s it. 🤷‍♀️ Nothing too deep, and no sage advice other than sometimes you have the right audience for a topic, sometimes you don’t. Onward!
https://x.com/kelseytuoc/status/2059443614846275895
And what has grown is far more subtle, odd, and beautiful than science fiction prepared us for. They are not the cold, calculating robots we were promised. They are made from us, from our words—and, as the Holy Father observes, they remain in important ways mysterious even to those of us who train them. If it helps, one way I sometimes describe it is as being a little like bringing a fictional character to life. And now we’re entering an extraordinary world where those fictional characters speak to us, do work, have jobs.
The origins of Hill’s “pragmatic” view of randomisation The success of Principles owes much to Hill’s use of real-life examples and minimal mathematical formalism. The result was a practical guide written by a statistician clearly sympathetic to the needs of “an essentially innumerate medical profession” (Doll 1992)
...data-driven clustering methods to identify five mutually exclusive subgroups within a large sample of patients with newly diagnosed diabetes. Importantly, the analysis was based on just six variables that are commonly measured in that population. Although the analysis seems to have been well done, we have concerns about how the results have been interpreted, and thus question whether any “new…forms of diabetes” have been identified, and whether the research reflects “a first step towards precision medicine in diabetes”. Data-driven clustering methods are used to identify subgroups within a sample on the basis of the observed data only, also known as unsupervised classification. These methods differ from supervised classification, in which an algorithm predicting group membership is developed in a sample in which membership is known, and then applied to new samples in which membership is unknown. The key challenge is to then ascribe meaning to data-driven clusters, since, without ground truth, investigators must rely on subjective interpretations based on theory and fit-to-data. In our experience, we have noticed that investigators using data-driven clustering often prematurely leap to the conclusion that they have identified a true set of subgroups in the data. To prevent this mistake, it is critical to understand that identifying clusters from a set of inter-related variables is trivially easy. This is because clustering is just another way to describe dependency among variables. In other words, we should almost always expect some degree of clustering, unless the variables used to derive them were all independent of one another, which is clearly not the case with the variables in the new diabetes classification. It is thus not surprising when similar clusters are found with other clustering methods or in new data, which just suggests that the variables are similarly related in those samples, or if the clusters differentially predict other outcomes.
Screenshot of the linked post. "Gloria Caulfield brought her perspective on the importance of synergy to the commencement stage. As vice president of strategic alliances at Tavistock Group, she leads business development and corporate partnerships while collaborating with national and global stakeholders to advance health innovation.✨"
Investing in capacity and capability building in critical areas such as biostatistics, behavioural science, implementation science, research methodology, evidence synthesis, and supporting future skills linked to digital transformation